Blockchain-Assisted Privacy And Security Enhancement In Federated Learning

Abstract

The rapid increase in road accidents caused by riders not wearing helmets and traffic rule violations has created the need for an intelligent automated monitoring system. This project, titled “Dual Detection of License Plates and Helmets Using an Optimized YOLO and Neural Network”, proposes a real-time traffic surveillance solution capable of simultaneously detecting motorcycle riders, identifying whether helmets are worn, and recognizing vehicle license plates.

The proposed system utilizes the YOLO (You Only Look Once) object detection algorithm due to its high speed and accuracy in real-time image processing. The optimized YOLO model is trained to detect helmets, motorcycles, and number plates from surveillance camera footage. After detecting the vehicle, Optical Character Recognition (OCR) techniques are applied to extract the license plate number. A Neural Network-based classification model enhances the accuracy of helmet detection under varying lighting, weather, and traffic conditions.

The system processes video frames continuously and identifies violations instantly. If a rider is detected without a helmet, the system captures the license plate and stores the violation details in the database for further action. Experimental results show high detection accuracy and fast processing speed, making the proposed system suitable for smart city traffic monitoring applications.

The proposed framework reduces manual traffic monitoring efforts, improves road safety enforcement, and provides a scalable intelligent transportation solution.